Our audit methodology is not opinion — it is peer-reviewed science. Every compliance criterion is grounded in published, reproducible research.
We do not guess. We prove. With 9 publications, our audit methodology is built on peer-reviewed data science. Frameworks currently submitted to IEEE, Springer, and Wiley. — Research Statement, Maqasid AI Labs
Our published work spans LLM safety, Islamic fintech, regulatory governance, and edge AI deployments.
Introduces a multi-layer filtering architecture grounded in Maqasid al-Shariah to detect and suppress religiously erroneous outputs generated by large language models in Islamic advisory contexts.
Presents a hardware-efficient transformer architecture optimized for real-time Shariah-compliance screening of financial transactions, validated across three Islamic banking datasets.
Proposes a counterfactual explainability framework aligned with Islamic ethical principles, enabling regulators and Shariah boards to audit AI decision-making processes transparently.
Develops a hardware-aware model compression pipeline that preserves Shariah-compliance properties during quantization and pruning for deployment on resource-constrained IoT devices.
Download the full bibliography or connect with our research profiles.